When Is Run-to-Failure a Sensible Maintenance Choice?

Bound the exception to one function and failure event

Run-to-failure means accepting a specified functional failure and responding afterward, instead of performing discretionary work intended to prevent that event. It does not mean abandoning the asset, ignoring damage or allowing a required task to expire. Define the exception more narrowly than “this machine is noncritical.”

Hypothetical worked scenario

Display VD-10 shows decorative welcome and factory-history images in a staffed visitor lobby. Assume a qualified site review confirms it has no emergency-direction, access, customer-check-in, production-control or quality function. These circumstances and every record below are invented, not a client assessment.

Mounting integrity, electrical protection and any mandatory or applicable manufacturer maintenance remain outside this exception. Any safety function or required task in the proposed scope excludes it from this route. Damage, overheating or other unsafe conditions follow existing qualified procedures, not the decorative-picture exception.

A blank screen alone does not identify a failed panel. The responsible reviewer must confirm that the event fits the approved boundary before applying the exception.

Find the people and systems that depend on it

Criticality describes consequence in context. An inexpensive display could also carry evacuation instructions, an access credential or the only available process-status view. Its name in an asset list would not reveal those dependencies. Ask the people who use it and inspect the actual content and connections.

For the hypothetical proposal, the content owner confirms that every displayed item is optional decoration, and the technical reviewer checks that losing the picture does not disable another service. A future change to emergency messaging would invalidate that basis, even if the hardware stayed identical.

Do not label a whole class of monitors, lights or small motors noncritical. Their assigned functions, location and connected systems determine the consequence of losing them. A classification copied from another room is a lead to investigate, not an approval.

Describe the loss the business is prepared to accept

Separate a stopped component from its business consequence. For VD-10, the proposed accepted loss is an absent welcome loop during the stated period. It is not an agreement to lose visitor registration, safety information or any service someone later chooses to display.

FHWA's lifecycle-planning primer says a reactive approach should consider repair time and the impact of downtime. Its examples concern transportation assets; they do not classify this factory display or justify accepting unknown consequences.

The business owner must state what interruption is acceptable and who notices it. If the next visitor campaign requires continuous images, that is a different service commitment. Do not preserve the exception by silently redefining the promised service after a failure.

EX-10 cannot be signed while the acceptable interruption and staffed response coverage remain unspecified. The readiness demonstration's duration is not an outage limit or a substitute for that business decision.

Low criticality also does not require buying a standby. If the decorative function can be absent without a substitute, record that choice openly. If the proposal depends on rapid substitution, readiness becomes part of its evidence.

Witness recovery rather than counting spare boxes

A spare display is useful only if it fits, works with the intended media source and can be retrieved and installed by authorized people. Include access, tools, mounting compatibility, configuration, content rights and the final picture check. A sealed carton does not establish any of those facts.

In the invented RD-10 rehearsal, retrieval takes 15 minutes and compatible setup plus verification takes 30 minutes: 15 + 30 = 45 minutes. The facilities lead witnesses it under the specified staffed conditions. This is hypothetical demonstration data, not an actual result or a future restoration guarantee.

Account separately for waiting before work starts. An unavailable technician or locked store can make the interruption much longer than the demonstrated task. Unknown access or recovery time blocks the proposal. If a spare is borrowed, reconfigured or used elsewhere, its earlier demonstration no longer establishes current readiness.

This spare provides a restoration path, not uninterrupted redundancy. A claimed standby function needs demonstrated compatibility, access and readiness in its intended role. Merely possessing another screen does not show that the required service remains available after failure.

Compare the full effort with realistic alternatives

Compare discretionary inspection, preventive replacement and monitoring with the consequences and recovery work being accepted. Include planning, spare storage, configuration upkeep, access and disposal. Run-to-failure is not a zero-cost option, and inspection is not free merely because someone already walks past the asset.

GSA's life-cycle guidance includes purchase, installation, operation, maintenance, training, removal and disposal in financial costs. Applying that perspective here means looking beyond the display's price; it does not establish which maintenance choice is cheapest.

Use attributable local estimates rather than invented failure rates or savings. Where costs or interruption impacts cannot be bounded, say so. Management can decide whether further evidence is proportionate, but a cheap component does not erase an unknown dependency or a required task.

Record who can approve exactly what

EX-10 belongs in the maintenance register as the system of record, linked to the dependency assessment, requirements review, RD-10 and alternatives. Keep proposed and approved status distinct. The facilities manager coordinates evidence; qualified technical reviewers establish the boundaries; authorized site management decides whether to accept the specified business consequence.

A blank approval field means pending, not consent. Any unresolved requirement, hidden dependency or recovery assumption keeps the exception unavailable. Current approved requirements remain in force. A new exception cannot authorize people to bypass manufacturer instructions or applicable legal and safety obligations.

Enter explicit effective and expiry dates, response coverage and the person receiving a failure report. A campaign description alone is insufficient. EX-10 stays pending here because those real-site commitments and signatures have not been supplied.

Approval of a maintenance policy also does not authorize physical replacement. Qualified people still follow the applicable work and return-to-service processes. Conventional software can record status and expiry; AI can summarize supporting evidence, but neither grants the exception.

Withdraw the exception when its basis changes

Set an expiry and event-driven review triggers, rather than treating the first assessment as permanent. After a failure, record the actual lost function, interruption, recovery work and unexpected effects. Compare those observations with EX-10's accepted scope before considering renewal.

A tabletop or witnessed demonstration supports only the conditions examined; it cannot guarantee every future recovery. The practical outcome is a reviewable decision with clear limits and a way to withdraw it, not a general claim that small assets should be allowed to fail.